Study On Methodology Of Variable Data Measurement System Capability And Assessment

In this dissertation, capability evaluation of attribute data measurement system, including non-destructive and destructive measurement systems, are under research in order to evaluate a broader measurement system.In this dissertation, fundamental characteristics and taxonomy of a measurement system are reviewed and then the evaluation of capability of non-destructive measurement system under typical circumstance and other types of conditions are carefully studied with ANOVA.Destructive measurement systems are clearly defined and classified, then according to the different systems'characteristics, five analysis methods are proposed: the homogeneous sampling, the alternative sampling, non-destructive measurement system alternative sampling, destructive measurement system alternative sampling, and time sequence sampling, which offer solutions and methodology for the analysis of destructive measurement systems.Basically, measurement system analysis is a multivariate analysis method, the rationality and analytical strategy will both affect the reliability of the final results. The number of measurers, samples and times of repeating are discussed to settle down a general alternative standard; at the same time, this dissertation offers testing and analyzing influence factors of the measurement system analysis model, using Residual analysis. What's more, a robust regression method M-estimators is adopted to eliminate the impact of the anomaly data.In order to get a reasonable judgment of the assessment result of the measurement system, on the research of system error's impact on the capacity of the process, this dissertation talks about the evaluation criteria of the capacity of the measurement system, when there are different requirements on the capacity of the process and error rate. Thus the evaluation criteria of the measurement system based on the capacity of the process is proposed.Last, in the field research of company X, a specific experiment is designed for the assessment of the capacity of the destructive measurement system in the company, and then the data of the experiment is analyzed and evaluated , the fitness of the model is checked, and abnormal values are analyzed. Therefore, we obtain the assessment of the measurement system, which is an effective way.